My R Timeline

• Digital marketing data since 2007
• useR since 2012 - Motive: how to use all this web data?
• Google Developer Expert - Google Analytics & Google Cloud
• Part of cloudyr group (AWS/Azure/GCP R packages for the cloud) https://cloudyr.github.io/
• Now: Data Engineer @ IIH Nordic

GA Effect

googleAuthRverse

• searchConsoleR
• googleAnalyticsR
• googleAuthR -> gargle
• googleComputeEngineR (cloudyr)
  
• googleCloudStorageR (cloudyr)
• bigQueryR (cloudyr)
• googleLanguageR (rOpenSci)
• googleCloudRunner (New!)

Slack: #googleAuthRverse

Agenda today

• Abstracting R applications into the Cloud using Docker (10min)
• Demo on what that abstraction offers (10 min)
• My evolving mindset for using R/Docker/Cloud (10 min)
• Any questions? (10 min)

What is..?

Docker - a container system for building and sharing applications

Cloud - computing delivered via the internet, not locally

What does serverless offer?

Serverless - cloud services that often use containers to host applications without configuring servers

• Focus on code, not dev-ops
• Scale from 0 to billions
• Reliability and security
• Abstraction

Climbing up the pyramid - R

Climbing up the pyramid - Cloud

The keystone

• R - abstraction of R environments
• Cloud - run any code on cloud systems

Docker + R = R in Production

• Flexible No need to ask IT to install R places, use docker run; across cloud platforms; ascendent tech

• Version controlled No worries new package releases will break code

• Scalable Run multiple Docker containers at once, fits into event-driven, stateless serverless future

Docker levels the playing ground between languages

Docker levels the playing ground between languages

Dockerfiles from The Rocker Project

https://www.rocker-project.org/

Maintain useful R images

• rocker/r-ver
• rocker/rstudio
• rocker/tidyverse
• rocker/shiny
• rocker/ml-gpu

Thanks to Rocker Team

Dockerfiles

FROM rocker/tidyverse:3.6.0
MAINTAINER Mark Edmondson (r@sunholo.com)

# install R package dependencies
RUN apt-get update && apt-get install -y \
    libssl-dev 

## Install packages from CRAN
RUN install2.r --error \ 
    -r 'http://cran.rstudio.com' \
    googleAuthR \ 
    googleComputeEngineR \ 
    googleAnalyticsR \ 
    searchConsoleR \ 
    googleCloudStorageR \
    bigQueryR \ 
    ## install Github packages
    && installGithub.r MarkEdmondson1234/youtubeAnalyticsR \
    ## clean up
    && rm -rf /tmp/downloaded_packages/ /tmp/*.rds \

Schedule an R script in the Cloud

Demo an Rmd script pulling data from googlesheets

Create an R API that scales from 0 to 1 billion

Demo a plumber script with Cloud Pub/Sub

Build a website for your package

Demo cloudbuild.yml building website via pkgdown

googleCloudRunner - Use Cases

install.packages("googleCloudRunner")
library(googleCloudRunner)

As easy as possible enabling of R use cases in the Cloud

• Scheduled R scripts
• Long-running R scripts
• Scale to 0 R APIs
• Continuous development

A use anywhere R/RStudio

• Configure RStudio Server just like home…?
• Doesn’t use the cloud to full potential

library(googleComputeEngineR)
gce_vm(template = "rstudio", name = "my-rstudio-server",
       username = "mark", password = "r2020",
       predefined_type = "n1-highmem-2")

A tailored R environment per task

• Workshops with material pre-loaded
• ML vs googleauthrverse vs shiny vs …

library(googleComputeEngineR)
gce_vm(name = "deeplearning-r", 
       template = "rstudio-gpu", 
       username = "mark", password = "mark123", 
       acceleratorType = "nvidia-tesla-k80")

Scaling up R environments

• library(future) for cluster parallel work
• Run many of the same R environments at a time
• R APIs, Shiny Apps

library(future)
library(googleComputeEngineR)

vms <- gce_vm_cluster()
plan(cluster, workers = as.cluster(vms))

f <- function(my_data, args){
   ## ....expensive...computations
   result
}

result %<-% f(my_data) 

Separation of code and data (and config)

• Good data science principle applicable in good cloud setups
• The right service for the right job - BigQuery, Cloud Storage

Kubernetes

Docker image originally for GCE, deployed to Kubernetes:

kubectl run shiny1 \
  --image gcr.io/gcer-public/shiny-googleauthrdemo:prod \
  --port 3838

kubectl expose deployment shiny1 \
  --target-port=3838  --type=NodePort

https://code.markedmondson.me/r-on-kubernetes-serverless-shiny-r-apis-and-scheduled-scripts/

Shiny apps on k8s

Batch jobs

Cloud Build (like the demo earlier - build Rmd on a schedule)

library(googleCloudRunner)
cr_deploy_packagetests()

R APIs

Cloud Run (like the demo earlier - trigger R code from events)

Adapt plumber API for the model

library(googleAnalyticsR)

#' Return output data from the ga_time_normalised ga_model
#' @param viewID The viewID for Google Analytics
#' @get /data
function(viewID=""){
  model <- ga_time_normalised(viewID)
  model$output
}

#' Plot out data from the ga_time_normalised ga_model
#' @param viewID The viewID for Google Analytics
#' @get /plot
#' @serializer htmlwidget
function(viewID=""){
  model <- ga_time_normalised(viewID)
  model$plot
}

htmlwidgets

@serialiser htmlwidgets is great

http://gallery.htmlwidgets.org/

Separation of code opens up possibilities

• Docker on VM - GPU support, ML dev work
• Kubernetes - Shiny apps
• Cloud Build - batch jobs
• Cloud Run - R APIs, event driven workflows

Take-aways

• Abstracting R environments using Docker opens up horizons on what R can do
• Abstraction offers ways to make hard things easier for common R tasks
• Separation of R environment code, data and config makes applications very transferable to the best tool for the job

Gratitude

• Thank you for listening
• Thanks to Anne Petersen for inviting me
• Thanks to R Core team for Rv1.0.0 and beyond
• Thanks to RStudio for all their cool things
• Thanks again to Rocker
• Thanks to Google for Developer Expert programme

Say hello

• @HoloMarkeD
• https://code.markedmondson.me
• Contact us at IIH Nordic to build scale for you mark@iihnordic.com
• Any questions now?